Navigating the AI Regulation Landscape

At Delphi, we’ve been closely following the recent developments in AI regulations, particularly the European Union's landmark AI Act. This legislation, poised to become the world's first comprehensive legal framework for trustworthy AI, presents both challenges and opportunities for companies like Delphi.

At the heart of Delphi's mission is the empowerment of experts, coaches, and creators to share their knowledge and expertise with the world through digital clones. However, as AI continues to advance, the concerns around deception, privacy, and identity have rightfully come to the forefront.

Addressing Deception

Delphi takes a proactive stance against the spread of misinformation. Our digital clones are programmed to always admit when they are, in fact, clones. They are not designed to deceive or mislead, but rather to serve as educational tools that can scale access to valuable expertise. Each clone's responses are grounded in factual information, with proper citations to ensure transparency. Additionally, our clones are equipped with the ability to acknowledge the limits of their knowledge, fostering an open and trustworthy dialogue.

Without these safeguards, digital clones could easily be used to spread misinformation and deceive users, undermining the trust and value they aim to provide. We've already seen bad actors use sophisticated deep fakes to create false videos of political candidates saying things they never said, in order to mislead and manipulate voters. By ensuring our clones are transparent about their nature, we empower creators to build meaningful connections with their audiences while maintaining a high standard of integrity and authenticity.

Safeguarding Privacy

Protecting the privacy of our users is of utmost importance to Delphi. We empower clone owners with complete control over their data and who has access to their digital counterparts. Upon request, we can even delete all data associated with a user, ensuring that their personal information is not shared or misused.

In contrast, a lack of privacy safeguards could lead to the exploitation of user data. At Delphi, we do not use any creator's data to train some underlying model - we think creators should always get credit for their data and should benefit from any public use of it. By putting users in control of their data, Delphi demonstrates its commitment to respecting the privacy and autonomy of creators and their audiences.

Safeguarding Identity

Delphi also places a strong emphasis on identity verification and protection. We only allow individuals to create digital clones of themselves or those they have explicit written permission to represent. During onboarding, we require users to submit a picture of themselves holding their photo ID. We currently manually verify each and every user that signs up to create a clone. We are constantly looking for better & more scalable ways to verify identity.

Failure to abide by this rule will result in a permanent ban from our platform (see our Terms of Use for more info). Furthermore, our "Delphi Immortal" tier enables users to license their digital likeness and guard against unauthorized use, ensuring that their identity remains firmly under their control.

Without these identity safeguards, digital clones could be used to impersonate individuals without their consent. Outside of Delphi, we've seen cases of bad actors using cloning technologies to impersonate family members and gain access to their bank accounts, or to run misinformation campaigns that damage someone's reputation. These abuses can lead to fraud, financial loss, and a deep breakdown of trust in the technology. By ensuring that users have full control over their digital identities, Delphi empowers creators to share their expertise while maintaining their personal and professional integrity.

Looking Ahead

As we navigate the evolving landscape of AI regulations, Delphi is committed to staying ahead of the curve. All responsible AI founders and companies should be putting in their own safeguards and engaging with policymakers, industry partners, and legal experts to ensure that their platforms and practices align with the principles of the EU AI Act and other emerging regulatory frameworks. We are also actively trying to innovate on ways to prove the origin of data (text, voice, and video) that our users train their clones on.

By embracing transparency, privacy, and identity safeguards, Delphi aims to set a new standard for the responsible development and deployment of digital cloning technologies. I am confident that our proactive approach to addressing the concerns around AI will not only future-proof our company but also contribute to the broader conversation on building a more trustworthy and ethical AI ecosystem.